e. WRF-VAR Diagnostics
WRF-Var produces a
number of diagnostic files that contain useful information on how the data
assimilation has performed. This section will introduce you to some of these
files, and what to look for.
By this stage of the online tutorial you
have successfully compiled and run WRF-Var. If not,
you should return to the WRF-Var Tutorial Page to try these out.
i. Which are the important diagnostic to look for?
Having run WRF-Var,
it is important to check a number of output files to see if the assimilation
appears sensible. Change directory to where you ran this case:
cd
ls -l
You will see something like the following:
-rw-r--r-- 1 rizvi
-rw-r--r-- 1 rizvi
-rw-r--r-- 1 rizvi
-rw-r--r-- 1 rizvi
-rw-r--r-- 1 rizvi
-rw-r--r-- 1 rizvi
drwxr-xr-x 2 rizvi
-rw-r--r-- 1 rizvi
drwxr-xr-x 2 rizvi
drwxr-xr-x 3 rizvi
Content of some useful diagnostic files are
as follows:
cost_fn and grad_fn:
Note: Make sure that you removed first two records (header) in cost_fn and
grad_fn. Also you will need to specify
the directory name for these two files.
gts_omb_oma: It holds (in ASCII format) information of each type
of observations, like its value, quality,
namelist.input: WRF-Var input namelist file. It displays all the non-default options
which user defined. If any of your namelist defined
options did not appear in this file, you may like to check its name and match
it with the “WRFV3/Registry/Registry.wrfvar”.
namelist.output: Consolidated list of all the namelist
options used.
rsl: Directory containing information of standard WRF-Var output from individual processors. It contains host of
information on number of observations, minimization, timings, etc. Additional
diagnostics may be printed in these files by including various “print” WRF-Var namelist options. To learn
more about these additional “print” options, search “print_” string in
“WRFV3/Registry/Registry.wrfvar”.
statistics: Textfile containing OMB (OI), OMA (OA) statistics (minimum,
maximum, mean and standard deviation) for each observation type and variable.
This information is very useful in diagnosing how WRF-Var
has used different components of the observing system. Also contained are the
analysis minus background (A-B) statistics, i.e., statistics of the analysis
increments for each model variable at each model level. This information is
very useful in checking the range of analysis increment values found in the
analysis, and where they are in the WRF-model grid space.
ii.
Where is the final analysis file?
The final analysis file resides in $FC_DIR/2007010200/wrfinput_d01.
It is in WRF (NetCDF) format.
You need to specify the analsyis_file
As an example, if you are aiming to display the
wind U-component field of the analysis at level 18, execute following command
after modifying the script “WRFV3/var/graphcs/ncl/WRF-Var_plot.ncl”
ncl
WRF-Var_plot.ncl
It will display like:
You may like to change the variable name,
level, etc., in this script to display the variable of your choice at the
desired sigma level.
Take time to look through the textual output
files to ensure you understand how WRF-Var has
performed. For example,
1. How closely has WRF-Var
fitted individual observation types? Look at the statistics file to
compare the O-B and O-A statistics.
2. How big are the analysis increments? Again, look in
the statistics file to see minimum/maximum values of A-B for each
variable at various levels. It will give you a feel of the impact of input
observation data you assimilated via WRF-Var by
modifying the input analysis first guess.
3. How long did WRF-Var take
to converge? Does it really converge? You will get the answers to all these
questions by looking into "rsl" files, as
it indicates the number of iterations taken by WRF-Var
to converge. If this is the same as the maximum number of iterations specified
in the namelist (NTMAX) or its default value (=200)
set in WRFV3/Registry/Registry.wrfvar”, then it means that the analysis
solution did not converge. If so, so you may like to increase the value of “NTMAX”
and re-run your case to ensure that the convergence is achieved.
iii. Plotting WRF-Var
analysis increments
A good visual way of seeing the impact of
assimilation of observations is to plot the analysis increments (i.e., analysis
minus first guess difference). There are many different graphics packages (e.g.,
RIP, NCL, GRADS etc) that can do this. The plot of level 18 theta increments
below was produced using a particular NCL script. This script is located at
“WRFV3/var/graphcs/ncl/WRF-Var_plot.ncl”
You need to modify this script to fix the
full path for first_guess and analysis files. You may
also like to modify the display level by setting “kl”
and the name of the variable to display by setting “var”.
Further details are given in this script.
If you are aiming to display increment of
potential temperature at level 18, after modifying
“WRFV3/var/graphcs/ncl/WRF-Var_plot.ncl” suitably, when you execute the
following command from“WRFV3/var/graphics/ncl”.
ncl WRF-Var_plot.ncl
The plot created will looks as follows:
Note: Higher the analysis increments, more
is the data impact in that region.
iv. What next?
OK, you have run WRF-Var,
checked out the diagnostics and are confident things are OK. Before running a
forecast, you must first modify the tendencies within the lateral boundary
condition files to be consistent with the new WRF-Var
initial conditions.
Trouble Shooting: